1/*
2*******************************************************************************
3*
4*   Copyright (C) 2009-2011, International Business Machines
5*   Corporation and others.  All Rights Reserved.
6*
7*******************************************************************************
8*   file name:  n2builder.h
9*   encoding:   US-ASCII
10*   tab size:   8 (not used)
11*   indentation:4
12*
13*   created on: 2009nov25
14*   created by: Markus W. Scherer
15*/
16
17#ifndef __N2BUILDER_H__
18#define __N2BUILDER_H__
19
20#include "unicode/utypes.h"
21
22#if !UCONFIG_NO_NORMALIZATION
23
24#include "unicode/errorcode.h"
25#include "unicode/unistr.h"
26#include "normalizer2impl.h"  // for IX_COUNT
27#include "toolutil.h"
28#include "utrie2.h"
29
30U_NAMESPACE_BEGIN
31
32extern UBool beVerbose, haveCopyright;
33
34struct Norm;
35
36class BuilderReorderingBuffer;
37class ExtraDataWriter;
38
39class Normalizer2DataBuilder {
40public:
41    Normalizer2DataBuilder(UErrorCode &errorCode);
42    ~Normalizer2DataBuilder();
43
44    enum OverrideHandling {
45        OVERRIDE_NONE,
46        OVERRIDE_ANY,
47        OVERRIDE_PREVIOUS
48    };
49
50    void setOverrideHandling(OverrideHandling oh);
51
52    enum Optimization {
53        OPTIMIZE_NORMAL,
54        OPTIMIZE_FAST
55    };
56
57    void setOptimization(Optimization opt) { optimization=opt; }
58
59    void setCC(UChar32 c, uint8_t cc);
60    void setOneWayMapping(UChar32 c, const UnicodeString &m);
61    void setRoundTripMapping(UChar32 c, const UnicodeString &m);
62    void removeMapping(UChar32 c);
63
64    void setUnicodeVersion(const char *v);
65
66    void writeBinaryFile(const char *filename);
67
68private:
69    friend class CompositionBuilder;
70    friend class Decomposer;
71    friend class ExtraDataWriter;
72    friend class Norm16Writer;
73
74    // No copy constructor nor assignment operator.
75    Normalizer2DataBuilder(const Normalizer2DataBuilder &other);
76    Normalizer2DataBuilder &operator=(const Normalizer2DataBuilder &other);
77
78    Norm *allocNorm();
79    Norm *getNorm(UChar32 c);
80    Norm *createNorm(UChar32 c);
81    Norm *checkNormForMapping(Norm *p, UChar32 c);  // check for permitted overrides
82
83    const Norm &getNormRef(UChar32 c) const;
84    uint8_t getCC(UChar32 c) const;
85    UBool combinesWithCCBetween(const Norm &norm, uint8_t lowCC, uint8_t highCC) const;
86    UChar32 combine(const Norm &norm, UChar32 trail) const;
87
88    void addComposition(UChar32 start, UChar32 end, uint32_t value);
89    UBool decompose(UChar32 start, UChar32 end, uint32_t value);
90    void reorder(Norm *p, BuilderReorderingBuffer &buffer);
91    UBool hasNoCompBoundaryAfter(BuilderReorderingBuffer &buffer);
92    void setHangulData();
93    int32_t writeMapping(UChar32 c, const Norm *p, UnicodeString &dataString);
94    void writeCompositions(UChar32 c, const Norm *p, UnicodeString &dataString);
95    void writeExtraData(UChar32 c, uint32_t value, ExtraDataWriter &writer);
96    int32_t getCenterNoNoDelta() {
97        return indexes[Normalizer2Impl::IX_MIN_MAYBE_YES]-Normalizer2Impl::MAX_DELTA-1;
98    }
99    void writeNorm16(UChar32 start, UChar32 end, uint32_t value);
100    void processData();
101
102    UTrie2 *normTrie;
103    UToolMemory *normMem;
104    Norm *norms;
105
106    int32_t phase;
107    OverrideHandling overrideHandling;
108
109    Optimization optimization;
110
111    int32_t indexes[Normalizer2Impl::IX_COUNT];
112    UTrie2 *norm16Trie;
113    UnicodeString extraData;
114    uint8_t smallFCD[0x100];
115
116    UVersionInfo unicodeVersion;
117};
118
119U_NAMESPACE_END
120
121#endif // #if !UCONFIG_NO_NORMALIZATION
122
123#endif  // __N2BUILDER_H__
124